The research proposed here will examine the psychological mechanisms involved in the elimination of unwanted learned behaviors. Specifically, the behavioral phenomenon of extinction will be examined from the perspective of a multi-component model of Pavlovian learning. Its long-range goal is to identify conditions under which extinction treatments may be expected to have durable effects on responding, and to provide a more complete understanding of extinction processes. One basic premise is that extinction treatments may have distinct effects on different components of Pavlovian learning (e.g., its more cognitive or emotional/motivational components). The present research will investigate this by using methods that dissect learning into its separate cognitive and emotional/motivational components in order to separately examine extinction effects on these distinct components. The research approach will be to use a rodent model (Rattus norvegicus) performing in an appetitive learning paradigm (magazine approach conditioning) because this paradigm has been used successfully to study the multiple components of Pavlovian learning to be dissected here. The results will be relevant to mental health concerns because this research will establish when extinction treatments may result in durable response loss, and, more generally, will point to the need to look at extinction treatment effects on different (e.g., cogniive and emotional) aspects of learned behavior. The research will explore 4 specific aims. Based on preliminary findings, Specific Aim 1 will test the hypothesis that stimuli given limited training o that have been compromised in some fashion will be especially vulnerable to extinction. It is anticipated that cognitively based learning (i.e., control by sensory- specific associations) will e undermined by extinction applied to these vulnerable stimuli. Specific Aim 2 will test the hypothesis that extinction after limited training is durable. If so, reduced control by sensory-specific associations should fail to recover in various situations designed to promote such recovery (i.e., spontaneous recovery, reinstatement, renewal, and reversal learning). Specific Aim 3 will directly test whether extinction undermines emotionally based learning more rapidly than it does cognitively based learning. This result will point to the importance of monitoring separate components of learning in assessing extinction. Specific Aim 4 will begin an exploration of one basic mechanism proposed to explain extinction, negative prediction errors. This will be examined by exploring the hypothesis that prediction errors are coded in a general emotional value currency. The idea will be tested by combining appetitive and aversively trained cues during extinction to determine if their summative effects predict the overall impact of extinction. Overall, the results from these studies will have important theoretical and therapeutic implications by emphasizing the need to assess multiple response systems in extinction and by determining when extinction may or may not be expected to work.